Combination of annotation methods

Working directory should contain root atlas Seurat object, which is available on GEO under record GSE152766:

Root_Atlas.rds (or get it by running through notebook 2, 3, 4 & 5)

Working directory must contain subdirectories,supp_data

supp_data/ should contain the files, which are available on Github (github.com/Hsu-Che-Wei/COPILOT):

Protoplasting_DEgene_FC2_list.txt
color_scheme_at.RData
Curated_Markers.csv
escoring.curated.marker.anno.nsds5.csv (get it by running through notebook 3)
escoring.curated.marker.csv (get it by running through notebook 3)
ici_raw_20200414.csv.bz2 (get it by running through notebook 4)

rm(list=ls())
# Set the working directory to where folders named after the samples are located. 
# The folder contains spliced.mtx, unspliced.mtx, barcodes and gene id files, and json files produced by scKB that documents the sequencing stats. 
setwd("/scratch/AG_Ohler/CheWei/scKB")

# Load libraries
suppressMessages(library(Seurat))
suppressMessages(library(scales))
suppressMessages(library(data.table))
suppressMessages(library(tidyverse))
suppressMessages(library(ggplot2))

sessionInfo()

R version 3.6.1 (2019-07-05)
Platform: x86_64-conda_cos6-linux-gnu (64-bit)
Running under: CentOS Linux 7 (Core)

Matrix products: default
BLAS/LAPACK: /fast/home/c/chsu/anaconda3/envs/seu314/lib/libopenblasp-r0.3.9.so

locale:
 [1] LC_CTYPE=en_US.utf-8       LC_NUMERIC=C              
 [3] LC_TIME=en_US.utf-8        LC_COLLATE=en_US.utf-8    
 [5] LC_MONETARY=en_US.utf-8    LC_MESSAGES=en_US.utf-8   
 [7] LC_PAPER=en_US.utf-8       LC_NAME=C                 
 [9] LC_ADDRESS=C               LC_TELEPHONE=C            
[11] LC_MEASUREMENT=en_US.utf-8 LC_IDENTIFICATION=C       

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
 [1] forcats_0.5.0     stringr_1.4.0     dplyr_1.0.0       purrr_0.3.4      
 [5] readr_1.3.1       tidyr_1.1.0       tibble_3.0.1      ggplot2_3.3.2    
 [9] tidyverse_1.3.0   data.table_1.12.8 scales_1.1.1      Seurat_3.1.5     

loaded via a namespace (and not attached):
 [1] nlme_3.1-148       tsne_0.1-3         fs_1.4.1           lubridate_1.7.9   
 [5] RcppAnnoy_0.0.16   RColorBrewer_1.1-2 httr_1.4.1         repr_1.1.0        
 [9] sctransform_0.2.1  tools_3.6.1        backports_1.1.8    R6_2.4.1          
[13] irlba_2.3.3        KernSmooth_2.23-17 uwot_0.1.8         DBI_1.1.0         
[17] lazyeval_0.2.2     colorspace_1.4-1   withr_2.2.0        tidyselect_1.1.0  
[21] gridExtra_2.3      compiler_3.6.1     cli_2.0.2          rvest_0.3.5       
[25] xml2_1.3.2         plotly_4.9.2.1     lmtest_0.9-37      ggridges_0.5.2    
[29] pbapply_1.4-2      rappdirs_0.3.1     pbdZMQ_0.3-3       digest_0.6.25     
[33] base64enc_0.1-3    pkgconfig_2.0.3    htmltools_0.5.0    dbplyr_1.4.4      
[37] readxl_1.3.1       htmlwidgets_1.5.1  rlang_0.4.6        rstudioapi_0.11   
[41] generics_0.0.2     zoo_1.8-8          jsonlite_1.7.0     ica_1.0-2         
[45] magrittr_1.5       patchwork_1.0.1    Matrix_1.2-18      fansi_0.4.1       
[49] Rcpp_1.0.4.6       IRkernel_1.1       munsell_0.5.0      ape_5.4           
[53] reticulate_1.16    lifecycle_0.2.0    stringi_1.4.6      MASS_7.3-51.6     
[57] Rtsne_0.15         plyr_1.8.6         blob_1.2.1         grid_3.6.1        
[61] parallel_3.6.1     listenv_0.8.0      ggrepel_0.8.2      crayon_1.3.4      
[65] lattice_0.20-41    haven_2.3.1        IRdisplay_0.7.0    cowplot_1.0.0     
[69] splines_3.6.1      hms_0.5.3          pillar_1.4.4       igraph_1.2.5      
[73] uuid_0.1-4         future.apply_1.5.0 reshape2_1.4.4     codetools_0.2-16  
[77] leiden_0.3.3       reprex_0.3.0       glue_1.4.1         evaluate_0.14     
[81] modelr_0.1.8       png_0.1-7          vctrs_0.3.1        cellranger_1.1.0  
[85] gtable_0.3.0       RANN_2.6.1         future_1.17.0      assertthat_0.2.1  
[89] rsvd_1.0.3         broom_0.5.6        survival_3.2-3     viridisLite_0.3.0 
[93] cluster_2.1.0      globals_0.12.5     fitdistrplus_1.1-1 ellipsis_0.3.1    
[97] ROCR_1.0-11       

1. Load data

# Load unwanted genes, which are genes being induced during protoplasting in our case
pp.genes <- as.character(read.table("./supp_data/Protoplasting_DEgene_FC2_list.txt", header=F)$V1)

# Load the color scheme for time and cell type annotation
load("./supp_data/color_scheme_at.RData")

# Load known markers, including both wet-lab validated and in-silico-derived 
known.good.markers <- read.csv("./supp_data/Curated_Markers.csv", header=T)  

# Load escoring output
escore <- read.csv("./supp_data/escoring.curated.marker.anno.nsds5.csv", header=T)
escore <- escore +1
es <- fread("./supp_data/escoring.curated.marker.csv", header = T)

# Load ICI output
ici_raw <- fread("./supp_data/ici_raw_20200414.csv.bz2")

# Read in the atlas
rc.integrated <- readRDS('./Root_Atlas.rds')

2. Process escoring output (marker-based annotation)

# A cell is annotated as the label which its marker has the highest escore 
call.index <- apply(es, 2, function(x){colnames(escore)[which(x==max(x))]});
pos.index <- apply(es, 2, function(x){if(max(x)<0){"negative"}else{"positive"}})

escore.anno <- c()
for (i in colnames(escore)){
  ea <- rep('NA',ncol(rc.integrated))  
  ea[escore[,i][!is.na(escore[,i])]] = as.character(known.good.markers$Celltype[match(i,known.good.markers$Locus)])
  escore.anno <- cbind(escore.anno,ea)
} 
spec.index <- apply(escore.anno,1,function(x){length(table(x[x!="NA"]))})

escore.anno <- rep('NA',ncol(rc.integrated)) 
for (i in 1:length(call.index)){ 
  if(pos.index[i] == "positive"){
    escore.anno[i] = as.character(known.good.markers$Celltype[match(call.index[i],known.good.markers$Locus)])
  }
} 

escore.anno[which(spec.index == 0)] <- 'NA'

# Store marker annotation results in atlas
rc.integrated$escore.anno <- escore.anno

3. Marker-based annotation

# Plot marker annotation
order <- c("Quiescent Center", "Columella", "Lateral Root Cap", "Atrichoblast", "Trichoblast", "Cortex", "Endodermis", "Pericycle", "Phloem", "Xylem", "Procambium", "NA")
palette <- c("#9400d3", "#5ab953", "#bfef45", "#008080", "#21B6A8", "#82b6ff", "#0000FF","#ff9900","#e6194b", "#9a6324", "#ffe119","#EEEEEE")
rc.integrated$escore.anno <- factor(rc.integrated$escore.anno , levels = order[sort(match(unique(rc.integrated$escore.anno),order))]) 
color <- palette[sort(match(unique(rc.integrated$escore.anno),order))]
options(repr.plot.width=12, repr.plot.height=10)
DimPlot(rc.integrated, reduction = "umap", group.by = "escore.anno", cols = color)+ggtitle("Escore nsds 5 Marker Annotation")

4. Process ICI output

# Select best ICI annotation based on adjusted-pvale and normalized ICI score 
ici_filtered <- ici_raw %>% group_by(Cell) %>% filter(p_adj == min(p_adj)) %>% filter(ici_score_norm == max(ici_score_norm)) %>% filter(p_val == max(p_val)) %>% filter(ici_score == max(ici_score)) 

# Select samples
ici_atlas <- ici_filtered %>% filter(Dataset == 'sc_12'|Dataset == 'sc_11'|Dataset == 'sc_30'|Dataset == 'sc_31'|Dataset == 'sc_37'|Dataset == 'sc_40'|Dataset == 'sc_51'|Dataset == 'sc_9_at'|Dataset == 'sc_10_at'|Dataset == 'sc_1'|Dataset == 'tnw1'|Dataset == 'tnw2'|Dataset == 'col0'|Dataset == 'pp1'|Dataset == 'dc1'|Dataset == 'dc2')

# Store ICI annotation results in atlas
ord <- unique(rc.integrated$orig.ident)
idx <- c()
for (i in 1:length(ord)){
    idx <- c(idx,gsub(paste0("_",i,"$"),"",gsub("^",paste0(ord[i],"_"),colnames(rc.integrated)[grep(paste0("_",i,"$"),colnames(rc.integrated))])))
}
rownames(ici_atlas) <- ici_atlas$Cell
rc.integrated$ici_celltype <- ici_atlas[idx,]$Cell_Type;
rc.integrated$ici_celltype_score <- ici_atlas[idx,]$ici_score;
rc.integrated$ici_celltype_p_val <- ici_atlas[idx,]$p_val;
rc.integrated$ici_celltype_score_norm <- ici_atlas[idx,]$ici_score_norm;
rc.integrated$ici_celltype_p_adj <- ici_atlas[idx,]$p_adj;
ici.order <- unique(rc.integrated$ici_celltype)

# Merge labels
rc.integrated$ici_celltype[which(rc.integrated$ici_celltype == "Columella_Stem")]="Columella";
rc.integrated$ici_celltype[which(rc.integrated$ici_celltype == "Ground_Cell_Stem")]="Unknown";
rc.integrated$ici_celltype[which(rc.integrated$ici_celltype == "Epidermis_LRC_Stem")]="Lateral_Root_Cap";
rc.integrated$ici_celltype[which(rc.integrated$ici_celltype == "Xylem_Stem")]="Meristematic_Xylem";
rc.integrated$ici_celltype[is.na(rc.integrated$ici_celltype)] = "Unknown"

5. ICI annotation

# Plot ICI annotation
ici.order <- unique(rc.integrated$ici_celltype)
ici.palette <- c('#C7EA46','#FF9900','#B67721','#21B6A8','#0082C8','#FE7F9C','#FF00FF','#0000FF','#008081','#FF0800','#C0C0C0','#5AB953','#9400D3','#7EF9FF')
rc.integrated$ici_celltype <- factor(rc.integrated$ici_celltype, levels = ici.order[sort(match(unique(rc.integrated$ici_celltype),ici.order))]) 
color.ici <- ici.palette[sort(match(unique(rc.integrated$ici_celltype),ici.order))]
options(repr.plot.width=10, repr.plot.height=8)
DimPlot(rc.integrated, reduction = "umap", group.by = "ici_celltype", cols = color.ici)+ggtitle("ici annotation")

Warning message:
“Setting row names on a tibble is deprecated.”

6. Correlation-based annotation

# Plot correlation-based annotation
color <- celltypepalette[sort(match(unique(rc.integrated$celltype.ID.P),celltypeorder))]
DimPlot(rc.integrated, reduction = "umap", group.by = "celltype.ID.P", cols = color)+ggtitle("RNA-seq correlation annotation")
rc.integrated$celltype.ID.P_significance <- rep("> 0.01", nrow(rc.integrated@meta.data))
rc.integrated$celltype.ID.P_significance[which(rc.integrated$celltype.pvalue.P <= 0.01)] = "<= 0.01"
rc.integrated$celltype.cor.P_significance <- rep(">= 0.6", nrow(rc.integrated@meta.data))
rc.integrated$celltype.cor.P_significance[which(rc.integrated$celltype.cor.P < 0.6)] = "< 0.6"
DimPlot(rc.integrated, reduction = "umap", group.by = "celltype.ID.P_significance", order = c("<= 0.01","> 0.01"),cols = c("#cccccc", "#ff4040"))
FeaturePlot(rc.integrated, reduction = "umap", features = "celltype.cor.P")
DimPlot(rc.integrated, reduction = "umap", group.by = "celltype.cor.P_significance", order = c(">= 0.6","< 0.6"),cols = c("#cccccc", "#008081"))
rc.integrated$Rad.ID.P <- factor(rc.integrated$Rad.ID.P, levels = radorder[sort(match(unique(rc.integrated$Rad.ID.P),radorder))])
color <- radpalette[sort(match(unique(rc.integrated$Rad.ID.P),radorder))]
DimPlot(rc.integrated, reduction = "umap", group.by = "Rad.ID.P", cols = color)+ggtitle("Microarray correlation annotation")
rc.integrated$Rad.ID.P_significance <- rep("> 0.01", nrow(rc.integrated@meta.data))
rc.integrated$Rad.ID.P_significance[which(rc.integrated$Rad.pvalue.P <= 0.01)] = "<= 0.01"
rc.integrated$Rad.cor.P_significance <- rep(">= 0.6", nrow(rc.integrated@meta.data))
rc.integrated$Rad.cor.P_significance[which(rc.integrated$Rad.cor.P < 0.6)] = "< 0.6"
DimPlot(rc.integrated, reduction = "umap", group.by = "Rad.ID.P_significance", order = c("<= 0.01","> 0.01"),cols = c("#cccccc", "#ff4040"))
FeaturePlot(rc.integrated, reduction = "umap", features = "Rad.cor.P")
DimPlot(rc.integrated, reduction = "umap", group.by = "Rad.cor.P_significance", order = c(">= 0.6","< 0.6"),cols = c("#cccccc", "#008081"))

7. Make initial consensus annotation

# For making consensus annotation, we only keep those with correlation coefficient > 0.6 and adjusted p-value/p-value < 0.01
# Correlation-based annotation
hc.rna.anno <- as.character(rc.integrated$celltype.ID.P);
hc.rna.anno[which(rc.integrated$celltype.pvalue.P > 0.01)] = "NA";
hc.rna.anno[which(rc.integrated$celltype.cor.P < 0.6)] = "NA";
hc.ma.anno <- as.character(rc.integrated$Rad.ID.P);
hc.ma.anno[which(rc.integrated$Rad.pvalue.P > 0.01)] = "NA";
hc.ma.anno[which(rc.integrated$Rad.cor.P < 0.6)] = "NA";
# ICI annotation
hc.ici.anno <- as.character(rc.integrated$ici_celltype);
hc.ici.anno[which(rc.integrated$ici_celltype_p_adj > 0.01)] = "NA";
# Marker-based annotation
hc.km.anno <- as.character(rc.integrated$escore.anno)

# Merge and unify labels of annotation
hc.km.anno[which(hc.km.anno == "Hair Cells")]="Trichoblast"
hc.km.anno[which(hc.km.anno == "Non-hair Cells")]="Atrichoblast"
hc.km.anno[which(hc.km.anno == "Protophloem sieve elements")]="Phloem"
hc.ma.anno[which(hc.ma.anno == "Hair Cell")]="Trichoblast"
hc.ma.anno[which(hc.ma.anno == "LRC")]="Lateral Root Cap"
hc.ma.anno[which(hc.ma.anno == "Mat.Xylem")]="Xylem"
hc.ma.anno[which(hc.ma.anno == "Meri.Xylem")]="Xylem"
hc.ma.anno[which(hc.ma.anno == "Non-Hair Cell")]="Atrichoblast"
hc.ma.anno[which(hc.ma.anno == "Phloem [S32]")]="Phloem"
hc.ma.anno[which(hc.ma.anno == "Phloem [SUC2]")]="Phloem"
hc.ma.anno[which(hc.ma.anno == "QC")]="Quiescent Center"
hc.ma.anno[which(hc.ma.anno == "Phloem Pole Pericycle")]="Pericycle"
hc.ma.anno[which(hc.ma.anno == "Xylem Pole Pericycle")]="Pericycle"
hc.rna.anno[which(hc.rna.anno == "columella")]="Columella"
hc.rna.anno[which(hc.rna.anno == "developing cortex")]="Cortex"
hc.rna.anno[which(hc.rna.anno == "developing xylem")]="Xylem"
hc.rna.anno[which(hc.rna.anno == "endodermis & QC cells")]="Endodermis"
hc.rna.anno[which(hc.rna.anno == "hair cells")]="Trichoblast"
hc.rna.anno[which(hc.rna.anno == "LRC & non-hair cells")]="Lateral Root Cap"
hc.rna.anno[which(hc.rna.anno == "matured cortex")]="Cortex"
hc.rna.anno[which(hc.rna.anno == "matured endodermis")]="Endodermis"
hc.rna.anno[which(hc.rna.anno == "matured xylem pole")]="Xylem"
hc.rna.anno[which(hc.rna.anno == "non-hair cells")]="Atrichoblast"
hc.rna.anno[which(hc.rna.anno == "phloem & companion cells")]="Phloem"
hc.rna.anno[which(hc.rna.anno == "phloem pole pericycle")]="Pericycle"
hc.rna.anno[which(hc.rna.anno == "protophloem & metaphloem")]="Phloem"
hc.rna.anno[which(hc.rna.anno == "QC cells")]="Quiescent Center"
hc.ici.anno[which(hc.ici.anno == "Mature_Cortex")]="Cortex"
hc.ici.anno[which(hc.ici.anno == "Phloem_Pole_Pericycle")]="Pericycle"
hc.ici.anno[which(hc.ici.anno == "Maturing_Xylem")]="Xylem"
hc.ici.anno[which(hc.ici.anno == "Developing_Cortex")]="Cortex"
hc.ici.anno[which(hc.ici.anno == "Phloem_Companion_Cells")]="Phloem"
hc.ici.anno[which(hc.ici.anno == "Meristematic_Xylem")]="Xylem"
hc.ici.anno[which(hc.ici.anno == "Quiescent_Center")]="Quiescent Center"
hc.ici.anno[which(hc.ici.anno == "Lateral_Root_Cap")]="Lateral Root Cap"
hc.ici.anno[which(hc.ici.anno == "Xylem_Pole_Pericycle")]="Pericycle"
hc.ici.anno[which(hc.ici.anno == "Phloem_Companion_Cells")]="Phloem"

# Merge annotations and metadata
cb.anno <- data.frame(hc.rna.anno=hc.rna.anno,hc.ma.anno=hc.ma.anno,hc.ici.anno=hc.ici.anno,hc.km.anno=hc.km.anno)
num.unique <- apply(cb.anno,1,function(x){length(unique(x[which(x!="NA")]))})
num.not.NA <- apply(cb.anno,1,function(x){length(which(x!="NA"))})
ici.not.NA <- apply(cb.anno,1,function(x){length(which(x[3]!="NA"))})
km.not.NA <- apply(cb.anno,1,function(x){length(which(x[4]!="NA"))})
cb.anno$num.unique <- num.unique 
cb.anno$num.not.NA <- num.not.NA 
cb.anno$ici.not.NA <- ici.not.NA
cb.anno$km.not.NA <- km.not.NA
cb.anno$Cell_ID <- colnames(rc.integrated)
cb.anno$order <- seq(1,ncol(rc.integrated))

# Voting, a cell needs at least two supports from different annotation methods in order to remain annotated. Marker-based annotation is given higher weight on final annotation decision
cb.anno.sub <- cb.anno %>% filter((num.not.NA >= 2 & num.unique <=2 & ici.not.NA==1) | (num.not.NA >= 2 & num.unique <=2 & km.not.NA==1)) 
cb.anno.sub$annotation <- apply(cb.anno.sub,1,function(x){ if (x[3]==x[4] & x[1] == x[2] & x[1]==x[3]){x[4]} else if (x[3]==x[4] & x[3]==x[1]){x[4]} else if (x[3]==x[4] & x[3]==x[2]){x[4]} else if (x[1]==x[2] & x[1]==x[4]){x[4]} else if (x[1]==x[2] & x[1]==x[3]){x[3]} else if ((x[3]=="Quiescent Center"| x[4] == "Quiescent Center") & x[5]==2){"Unknown"} else if ((x[3] == "Phloem" | x[4]=="Phloem") & x[5]==2){"Unknown"} else if ((x[3] == "Pericycle" | x[4]=="Pericycle") & x[5]==2){"Unknown"} else if ((x[3] == "Xylem" | x[4]=="Xylem") & x[5]==2){"Unknown"} else if ((x[3] == "Endodermis" | x[4]=="Endodermis") & x[5]==2){"Unknown"} else if ((x[3] == "Cortex" | x[4]=="Cortex") & x[5]==2){"Unknown"} else if (x[1]==x[2] & x[1]==x[3]){x[3]} else if (x[1]==x[4] & x[4] != "NA"){x[4]} else if (x[2]==x[4] & x[4] != "NA"){x[4]} else if (x[1]==x[3] & x[3] != "NA"){x[3]} else if (x[2]==x[3]  & x[3] != "NA"){x[3]} else if (x[4] == "NA"){x[3]} else {x[4]}})
rc.integrated$cb.anno <- rep("Unknown", ncol(rc.integrated))
rc.integrated$cb.anno[as.numeric(cb.anno.sub$order)] <- as.character(cb.anno.sub$annotation)

# Plot initial consensus annotation
order <- c("Quiescent Center", "Columella", "Lateral Root Cap", "Atrichoblast", "Trichoblast", "Cortex", "Endodermis", "Pericycle", "Phloem", "Xylem", "Procambium", "Unknown")
palette <- c("#9400d3", "#5ab953", "#bfef45", "#008080", "#21B6A8", "#82b6ff", "#0000FF","#ff9900","#e6194b", "#9a6324", "#ffe119","#EEEEEE")
rc.integrated$cb.anno <- factor(rc.integrated$cb.anno, levels = order[sort(match(unique(rc.integrated$cb.anno),order))]) 
color <- palette[sort(match(unique(rc.integrated$cb.anno),order))]
options(repr.plot.width=10, repr.plot.height=8)
DimPlot(rc.integrated, reduction = "umap", group.by = "cb.anno", cols = color)+ggtitle("consensus annotation")

8. Build new reference expression profiles based on initial consensus annotation

# Merge cell type consensus annotation with time zone correlation-based annotation
long.merge <- as.character(rc.integrated$Long.ID.P)
cb.anno.t <- data.frame(rna=as.character(rc.integrated$timezone.ID.P), ma=long.merge)
cb.anno.t$order <- seq(1,ncol(rc.integrated))
cb.anno.t$annotation <- apply(cb.anno.t,1,function(x){if(TRUE){x[2]}else{"Unknown"}})
cb.anno.t.sub <- cb.anno.t %>% filter(annotation!="Unknown")
cb.anno.m <- left_join(cb.anno.sub,cb.anno.t.sub,by="order")
cb.anno.m <- cb.anno.m %>% filter(annotation.x != "NA" & annotation.x != "Unknown" & annotation.y != "NA")

# Disect each cell type based on microarray time zone correlation-based annotation to have more grouping of cells/ to give finest resolution 
cb.anno.m$final.annotation <- paste(cb.anno.m$annotation.y, cb.anno.m$annotation.x, sep="-")

# Filter grouping that has too few cells
cb.anno.m$final.annotation[grep(paste(names(table(cb.anno.m$final.annotation)[which(table(cb.anno.m$final.annotation) <= 2)]),collapse = "|"),cb.anno.m$final.annotation)] = "Unknown"
cb.anno.m <- cb.anno.m %>% filter(final.annotation != "Unknown")

# 88 new reference expression profiles are going to be build
length(sort(table(cb.anno.m$final.annotation)))
rc.integrated$cb.anno <- rep("Unknown", ncol(rc.integrated))
rc.integrated$cb.anno[as.numeric(cb.anno.m$order)] <- as.character(cb.anno.m$final.annotation)

88

# Extract integrated (batch-corrected) expression matrix
afm <- rc.integrated@assays$integrated@data

# Pool (average) expression values of each grouping
new_ref <- matrix(nrow=nrow(afm), ncol = 0)
for (i in 1:88) {
m <- afm[,which(rc.integrated$cb.anno==as.character(unique(rc.integrated$cb.anno[which(rc.integrated$cb.anno!="Unknown")]))[i])]
new_ref <- cbind(new_ref, rowSums(m)/ncol(m))
}
colnames(new_ref) <- as.character(unique(rc.integrated$cb.anno[which(rc.integrated$cb.anno!="Unknown")]))
gene.var <- apply(new_ref,1,var)

# Select top 200 highly variable genes 
new_ref_sub <- new_ref[names(sort(gene.var,decreasing = TRUE)[1:200]),]

# Merge 88 newly-built reference with atlas

merge.rownames <- function (x,y){
  dat <- merge(x = x, y = y, by = "row.names")
  rownames(dat) <- dat$Row.names
  dat <- dat[,-1]
  return(dat)
}

nr <- Reduce(merge.rownames, list(new_ref_sub,afm)) 
nr <- as.matrix(nr)

colnames(nr)[1:89]

1. 'Meri-5-Atrichoblast'
2. 'Elong-7-Cortex'
3. 'Meri-3-Lateral Root Cap'
4. 'Meri-5-Lateral Root Cap'
5. 'Elong-8-Pericycle'
6. 'Elong-8-Xylem'
7. 'Elong-8-Endodermis'
8. 'Elong-8-Phloem'
9. 'Mat-10-Cortex'
10. 'Mat-11-Columella'
11. 'Mat-10-Atrichoblast'
12. 'Elong-8-Procambium'
13. 'Meri-6-Trichoblast'
14. 'Mat-10-Procambium'
15. 'Elong-7-Atrichoblast'
16. 'Meri-6-Cortex'
17. 'Mat-10-Trichoblast'
18. 'Elong-7-Endodermis'
19. 'Meri-1-Columella'
20. 'Elong-7-Trichoblast'
21. 'Elong-7-Procambium'
22. 'Mat-11-Trichoblast'
23. 'Mat-10-Xylem'
24. 'Meri-3-Atrichoblast'
25. 'Meri-3-Columella'
26. 'Meri-5-Columella'
27. 'Mat-10-Endodermis'
28. 'Meri-3-Pericycle'
29. 'Mat-11-Lateral Root Cap'
30. 'Mat-11-Atrichoblast'
31. 'Meri-5-Trichoblast'
32. 'Elong-7-Xylem'
33. 'Meri-2-Atrichoblast'
34. 'Mat-11-Cortex'
35. 'Mat-11-Pericycle'
36. 'Meri-6-Atrichoblast'
37. 'Elong-7-Lateral Root Cap'
38. 'Meri-3-Trichoblast'
39. 'Elong-8-Cortex'
40. 'Meri-2-Quiescent Center'
41. 'Mat-10-Pericycle'
42. 'Meri-5-Cortex'
43. 'Meri-1-Lateral Root Cap'
44. 'Elong-7-Pericycle'
45. 'Mat-9-Trichoblast'
46. 'Meri-2-Lateral Root Cap'
47. 'Meri-5-Phloem'
48. 'Meri-5-Xylem'
49. 'Mat-11-Phloem'
50. 'Meri-3-Cortex'
51. 'Mat-11-Xylem'
52. 'Mat-10-Lateral Root Cap'
53. 'Mat-11-Procambium'
54. 'Mat-10-Phloem'
55. 'Meri-3-Procambium'
56. 'Elong-7-Columella'
57. 'Elong-8-Trichoblast'
58. 'Meri-3-Xylem'
59. 'Meri-6-Phloem'
60. 'Meri-3-Quiescent Center'
61. 'Meri-2-Columella'
62. 'Elong-8-Atrichoblast'
63. 'Elong-7-Phloem'
64. 'Meri-3-Endodermis'
65. 'Mat-10-Columella'
66. 'Meri-1-Xylem'
67. 'Mat-11-Endodermis'
68. 'Mat-9-Endodermis'
69. 'Elong-8-Columella'
70. 'Meri-5-Pericycle'
71. 'Meri-4-Lateral Root Cap'
72. 'Meri-4-Xylem'
73. 'Meri-5-Endodermis'
74. 'Meri-4-Atrichoblast'
75. 'Meri-3-Phloem'
76. 'Mat-9-Cortex'
77. 'Meri-6-Xylem'
78. 'Meri-4-Columella'
79. 'Meri-6-Lateral Root Cap'
80. 'Meri-1-Atrichoblast'
81. 'Meri-5-Procambium'
82. 'Meri-6-Columella'
83. 'Mat-12-Lateral Root Cap'
84. 'Mat-9-Lateral Root Cap'
85. 'Columella-Lateral Root Cap'
86. 'Mat-9-Columella'
87. 'Mat-12-Phloem'
88. 'Meri-2-Procambium'
89. 'AAACCCAAGCCATTCA_1'

# Correlation-based annotation using 88 newly-built references
nr_label=colnames(new_ref)
nr_stat <- suppressWarnings(sapply(89:ncol(nr), function(i) sapply(1:88, function(j) cor.test(nr[,i],nr[,j],method = "pearson")[c(3,4)])))
nr_cor <- nr_stat[seq(2,nrow(nr_stat),2),]
nr_pvalue <- nr_stat[seq(1,nrow(nr_stat)-1,2),]                                                                                                        
nr_max <- sapply(1:(ncol(nr)-88), function(i) max(as.numeric(nr_cor[,i])))
nr_ident <- sapply(1:(ncol(nr)-88), function(i) nr_label[which(as.numeric(nr_cor[,i])==max(as.numeric(nr_cor[,i])))])
nr_maxp <- sapply(1:(ncol(nr)-88), function(i) as.numeric(nr_pvalue[,i])[which(as.numeric(nr_cor[,i])==max(as.numeric(nr_cor[,i])))])                     
names(nr_max) <- nr_ident
rc.integrated$final.anno.P <- as.character(nr_ident)
rc.integrated$final.anno.cor.P <- as.numeric(nr_max)
rc.integrated$final.anno.pvalue.P <- as.numeric(nr_maxp)
rc.integrated$final.anno.wo.unknown <-  gsub("^Columella-","",gsub("^Meri-.*-","",gsub("^Elong-.*-","",gsub("^Mat-.*-","",as.character(rc.integrated$final.anno.P)))))

# Plot new correlation-based annotation
rc.integrated$final.anno.wo.unknown <-  gsub("^Columella-","",gsub("^Meri-.*-","",gsub("^Elong-.*-","",gsub("^Mat-.*-","",as.character(rc.integrated$final.anno.P)))))
order <- c("Quiescent Center", "Columella", "Lateral Root Cap", "Atrichoblast", "Trichoblast", "Cortex", "Endodermis", "Pericycle", "Phloem", "Xylem", "Procambium", "Unknown")
palette <- c("#9400d3", "#5ab953", "#bfef45", "#008080", "#21B6A8", "#82b6ff", "#0000FF","#ff9900","#e6194b", "#9a6324", "#ffe119","#EEEEEE")
rc.integrated$final.anno.wo.unknown <- factor(rc.integrated$final.anno.wo.unknown, levels = order[sort(match(unique(rc.integrated$final.anno.wo.unknown),order))]) 
color <- palette[sort(match(unique(rc.integrated$final.anno.wo.unknown),order))]
options(repr.plot.width=12, repr.plot.height=10)
DimPlot(rc.integrated, reduction = "umap", group.by = "final.anno.wo.unknown", cols = color)+ggtitle("final annotation")

# Store index of QC
QC.idx <- which(rc.integrated$final.anno.wo.unknown=="Quiescent Center")

# Plot correlation coefficient
rc.integrated$final.anno.P_significance <- rep(">= 0.6", nrow(rc.integrated@meta.data))
rc.integrated$final.anno.P_significance[which(rc.integrated$final.anno.cor.P < 0.6)] = "< 0.6"
FeaturePlot(rc.integrated, reduction = "umap", features = "final.anno.cor.P")
DimPlot(rc.integrated, reduction = "umap", group.by = "final.anno.P_significance", order = c(">= 0.6","< 0.6"),cols = c("#cccccc", "#008081"))

9. Make second consensus annotation

# Make second consensus annotation by keeping annotation of those in agreement with marker-based annotation
cb.anno.df <- data.frame(order = seq(1,ncol(rc.integrated)),escore.anno = as.character(rc.integrated$escore.anno), final.anno.wo.unknown = as.character(rc.integrated$final.anno.wo.unknown))
cb.anno.df$selected <- apply(cb.anno.df, 1 ,function(x){if (x[2]==x[3]) {x[2]} else {"Unknown"}})

rc.integrated$cb.anno.2 <- cb.anno.df$selected

# Plot second consensus annotation
order <- c("Quiescent Center", "Columella", "Lateral Root Cap", "Atrichoblast", "Trichoblast", "Cortex", "Endodermis", "Pericycle", "Phloem", "Xylem", "Procambium", "Unknown")
palette <- c("#9400d3", "#5ab953", "#bfef45", "#008080", "#21B6A8", "#82b6ff", "#0000FF","#ff9900","#e6194b", "#9a6324", "#ffe119","#EEEEEE")
rc.integrated$cb.anno.2 <- factor(rc.integrated$cb.anno.2, levels = order[sort(match(unique(rc.integrated$cb.anno.2),order))]) 
color <- palette[sort(match(unique(rc.integrated$cb.anno.2),order))]
options(repr.plot.width=12, repr.plot.height=10)
DimPlot(rc.integrated, reduction = "umap", group.by = "cb.anno.2", cols = color)+ggtitle("cb.anno.2")

10. Derive time zone annotation on second consensus annotation by leveraging microarray, RNA-seq and CytoTRACE information

# Merge cell type consensus annotation with time zone correlation-based annotation
long.merge <- as.character(rc.integrated$Long.ID.P)
long.merge[grep("Meri",long.merge)]="Meristem";
long.merge[grep("Elong",long.merge)]="Elongation";
long.merge[grep("Mat",long.merge)]="Maturation";
long.merge[grep("Columella",long.merge)]="Maturation"
cb.anno.t <- data.frame(rna=as.character(rc.integrated$timezone.ID.P), ma=long.merge)
cb.anno.t$order <- seq(1,ncol(rc.integrated))

# Keep cell annotation of those having the same label of microarray and RNA-seq correlation-based time zone annotation 
cb.anno.t$annotation <- apply(cb.anno.t,1,function(x){if(x[1]==x[2]){x[2]}else{"Unknown"}})
cb.anno.t.sub <- cb.anno.t %>% filter(annotation!="Unknown")

cb.anno.m <- left_join(cb.anno.df,cb.anno.t.sub,by="order")

cb.anno.m <- cb.anno.m %>% filter(selected != "Unknown") %>% filter(annotation == "Meristem" | annotation == "Elongation" | annotation == "Maturation")

cb.anno.m$final.annotation <- paste(cb.anno.m$annotation, cb.anno.m$selected, sep="-")

sort(table(cb.anno.m$final.annotation))

Elongation-Quiescent Center         Meristem-Endodermis 
                          9                          79 
             Meristem-Xylem           Maturation-Phloem 
                         91                         171 
            Meristem-Phloem             Meristem-Cortex 
                        171                         179 
       Maturation-Pericycle         Meristem-Procambium 
                        250                         350 
         Meristem-Pericycle Elongation-Lateral Root Cap 
                        413                         456 
      Maturation-Procambium            Maturation-Xylem 
                        505                         553 
Maturation-Lateral Root Cap        Elongation-Columella 
                        560                         603 
  Meristem-Quiescent Center        Meristem-Trichoblast 
                       1088                        1512 
          Maturation-Cortex     Maturation-Atrichoblast 
                       2066                        2225 
           Elongation-Xylem       Meristem-Atrichoblast 
                       2288                        2619 
         Meristem-Columella       Maturation-Endodermis 
                       2865                        2921 
       Maturation-Columella      Maturation-Trichoblast 
                       2949                        3338 
     Elongation-Trichoblast           Elongation-Phloem 
                       3808                        3848 
    Elongation-Atrichoblast   Meristem-Lateral Root Cap 
                       4380                        5183 
          Elongation-Cortex       Elongation-Endodermis 
                       5421                        5609 
       Elongation-Pericycle       Elongation-Procambium 
                       6972                        7380 

# Filter grouping that has too few cells
cb.anno.m$final.annotation[grep(paste(names(table(cb.anno.m$final.annotation)[which(table(cb.anno.m$final.annotation) <= 20)]),collapse = "|"),cb.anno.m$final.annotation)] = "Unknown"
cb.anno.m <- cb.anno.m %>% filter(final.annotation != "Unknown")

# 31 new reference expression profiles 
length(sort(table(cb.anno.m$final.annotation)))
rc.integrated$cb.anno.t <- rep("Unknown", ncol(rc.integrated))
rc.integrated$cb.anno.t[as.numeric(cb.anno.m$order)] <- as.character(cb.anno.m$final.annotation)

31

# Load CytoTRACE results
cb.anno.m$CytoTRACE <- rc.integrated$CytoTRACE[cb.anno.m$order]

sub <- cb.anno.m %>% group_by(final.annotation) %>% arrange(final.annotation,CytoTRACE)

sub

A grouped_df: 70853 × 9

order	escore.anno	final.anno.wo.unknown	selected	rna	ma	annotation	final.annotation	CytoTRACE
<int>	<fct>	<fct>	<chr>	<fct>	<fct>	<chr>	<chr>	<dbl>
1784	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
13977	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
14649	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
15184	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
20937	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
21761	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
24329	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
27021	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
30075	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
37438	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
38357	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
41025	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
47348	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
50151	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
59324	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
60118	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
64649	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
66709	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
66725	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
83165	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
85367	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
85522	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
90950	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
92779	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
96570	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
96797	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.04704705
20429	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.05305305
28681	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.05305305
85610	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.05305305
88596	Atrichoblast	Atrichoblast	Atrichoblast	Elongation	Elongation	Elongation	Elongation-Atrichoblast	0.05305305
⋮	⋮	⋮	⋮	⋮	⋮	⋮	⋮	⋮
99388	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.6403070
101733	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.6403070
104002	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.6403070
109294	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.6403070
109896	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.6403070
4957	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.6966967
14506	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.6966967
14530	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.6966967
15053	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.6966967
16231	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.6966967
17455	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.6966967
17854	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.6966967
23958	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.6966967
30565	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.6966967
39484	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.6966967
41748	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.6966967
50337	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.6966967
56434	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.6966967
84601	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.6966967
89038	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.6966967
98966	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.6966967
92703	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.7097097
2931	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.7697698
26364	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.7697698
30484	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.7697698
38480	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.7697698
42281	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.7697698
57406	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.7697698
66554	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.7697698
96462	Xylem	Xylem	Xylem	Meristem	Meristem	Meristem	Meristem-Xylem	0.7697698

# Counsult CytoTRACE to produce finer groupping
sub1 <- sub %>%
  group_by(final.annotation) %>%
  summarise(median(CytoTRACE), n())
Group <- c()
for (i in 1:nrow(sub1)) {
  a <- rep(sub1$`median(CytoTRACE)`[i], sub1$`n()`[i])
  Group <- c(Group, a)
}
sub$Median_cycto <- Group
sub2 <- sub %>%
  mutate(Group = ifelse(test = CytoTRACE >= Median_cycto, yes = "G1",
                        no = "G2"))
Check <- sub2 %>%
  group_by(final.annotation, Group) %>%
  summarise(n())

sub2$cttz <- paste(sub2$final.annotation, sub2$Group, sep="-")

sort(table(sub2$cttz))

        Meristem-Endodermis-G2         Meristem-Endodermis-G1 
                            39                             40 
             Meristem-Xylem-G2              Meristem-Xylem-G1 
                            41                             50 
          Maturation-Phloem-G2             Meristem-Cortex-G2 
                            61                             76 
            Meristem-Phloem-G2             Meristem-Phloem-G1 
                            84                             87 
            Meristem-Cortex-G1           Maturation-Phloem-G1 
                           103                            110 
       Maturation-Pericycle-G2        Maturation-Pericycle-G1 
                           113                            137 
        Meristem-Procambium-G2         Meristem-Procambium-G1 
                           155                            195 
         Meristem-Pericycle-G2          Meristem-Pericycle-G1 
                           199                            214 
Elongation-Lateral Root Cap-G1 Elongation-Lateral Root Cap-G2 
                           228                            228 
      Maturation-Procambium-G2       Maturation-Procambium-G1 
                           251                            254 
           Maturation-Xylem-G2 Maturation-Lateral Root Cap-G1 
                           271                            280 
Maturation-Lateral Root Cap-G2            Maturation-Xylem-G1 
                           280                            282 
       Elongation-Columella-G2        Elongation-Columella-G1 
                           301                            302 
  Meristem-Quiescent Center-G2   Meristem-Quiescent Center-G1 
                           541                            547 
       Meristem-Trichoblast-G2        Meristem-Trichoblast-G1 
                           719                            793 
          Maturation-Cortex-G2           Maturation-Cortex-G1 
                          1024                           1042 
    Maturation-Atrichoblast-G2            Elongation-Xylem-G2 
                          1096                           1117 
    Maturation-Atrichoblast-G1            Elongation-Xylem-G1 
                          1129                           1171 
      Meristem-Atrichoblast-G2       Meristem-Atrichoblast-G1 
                          1285                           1334 
         Meristem-Columella-G2          Meristem-Columella-G1 
                          1431                           1434 
       Maturation-Columella-G2       Maturation-Endodermis-G2 
                          1449                           1459 
      Maturation-Endodermis-G1        Maturation-Columella-G1 
                          1462                           1500 
     Maturation-Trichoblast-G1      Maturation-Trichoblast-G2 
                          1669                           1669 
     Elongation-Trichoblast-G2           Elongation-Phloem-G2 
                          1874                           1922 
          Elongation-Phloem-G1      Elongation-Trichoblast-G1 
                          1926                           1934 
    Elongation-Atrichoblast-G1     Elongation-Atrichoblast-G2 
                          2190                           2190 
  Meristem-Lateral Root Cap-G2   Meristem-Lateral Root Cap-G1 
                          2571                           2612 
          Elongation-Cortex-G2           Elongation-Cortex-G1 
                          2708                           2713 
      Elongation-Endodermis-G2       Elongation-Endodermis-G1 
                          2785                           2824 
       Elongation-Pericycle-G2        Elongation-Pericycle-G1 
                          3457                           3515 
      Elongation-Procambium-G2       Elongation-Procambium-G1 
                          3673                           3707 

sub2 <- sub2 %>% filter(cttz != "Unknown")
sub2$cttz[grep("Meristem-Quiescent Center",sub2$cttz)] = "Meristem-Quiescent Center";
sub2$cttz[grep("Meristem-Lateral Root Cap",sub2$cttz)] = "Meristem-Lateral Root Cap";
sub2$cttz[grep("Elongation-Lateral Root Cap",sub2$cttz)] = "Elongation-Lateral Root Cap";
sub2$cttz[grep("Maturation-Lateral Root Cap",sub2$cttz)] = "Maturation-Lateral Root Cap";
sub2$cttz[grep("Meristem-Columella",sub2$cttz)] = "Meristem-Columella";
sub2$cttz[grep("Elongation-Columella",sub2$cttz)] = "Elongation-Columella";
sub2$cttz[grep("Maturation-Columella",sub2$cttz)] = "Maturation-Columella";

# 55 new reference expression profiles are going to be build
length(sort(table(sub2$cttz)))
rc.integrated$cb.anno.t <- rep("Unknown", ncol(rc.integrated))
rc.integrated$cb.anno.t[as.numeric(sub2$order)] <- as.character(sub2$cttz)

55

table(rc.integrated$cb.anno.t)

 Elongation-Atrichoblast-G1  Elongation-Atrichoblast-G2 
                       2190                        2190 
       Elongation-Columella        Elongation-Cortex-G1 
                        603                        2713 
       Elongation-Cortex-G2    Elongation-Endodermis-G1 
                       2708                        2824 
   Elongation-Endodermis-G2 Elongation-Lateral Root Cap 
                       2785                         456 
    Elongation-Pericycle-G1     Elongation-Pericycle-G2 
                       3515                        3457 
       Elongation-Phloem-G1        Elongation-Phloem-G2 
                       1926                        1922 
   Elongation-Procambium-G1    Elongation-Procambium-G2 
                       3707                        3673 
  Elongation-Trichoblast-G1   Elongation-Trichoblast-G2 
                       1934                        1874 
        Elongation-Xylem-G1         Elongation-Xylem-G2 
                       1171                        1117 
 Maturation-Atrichoblast-G1  Maturation-Atrichoblast-G2 
                       1129                        1096 
       Maturation-Columella        Maturation-Cortex-G1 
                       2949                        1042 
       Maturation-Cortex-G2    Maturation-Endodermis-G1 
                       1024                        1462 
   Maturation-Endodermis-G2 Maturation-Lateral Root Cap 
                       1459                         560 
    Maturation-Pericycle-G1     Maturation-Pericycle-G2 
                        137                         113 
       Maturation-Phloem-G1        Maturation-Phloem-G2 
                        110                          61 
   Maturation-Procambium-G1    Maturation-Procambium-G2 
                        254                         251 
  Maturation-Trichoblast-G1   Maturation-Trichoblast-G2 
                       1669                        1669 
        Maturation-Xylem-G1         Maturation-Xylem-G2 
                        282                         271 
   Meristem-Atrichoblast-G1    Meristem-Atrichoblast-G2 
                       1334                        1285 
         Meristem-Columella          Meristem-Cortex-G1 
                       2865                         103 
         Meristem-Cortex-G2      Meristem-Endodermis-G1 
                         76                          40 
     Meristem-Endodermis-G2   Meristem-Lateral Root Cap 
                         39                        5183 
      Meristem-Pericycle-G1       Meristem-Pericycle-G2 
                        214                         199 
         Meristem-Phloem-G1          Meristem-Phloem-G2 
                         87                          84 
     Meristem-Procambium-G1      Meristem-Procambium-G2 
                        195                         155 
  Meristem-Quiescent Center     Meristem-Trichoblast-G1 
                       1088                         793 
    Meristem-Trichoblast-G2           Meristem-Xylem-G1 
                        719                          50 
          Meristem-Xylem-G2                     Unknown 
                         41                       39574 

# Extract integrated (batch-corrected) expression matrix
afm <- rc.integrated@assays$integrated@data

# Pool (average) expression values of each grouping
new_ref <- matrix(nrow=nrow(afm), ncol = 0)
for (i in 1:55) {
m <- afm[,which(rc.integrated$cb.anno.t==as.character(unique(rc.integrated$cb.anno.t[which(rc.integrated$cb.anno.t!="Unknown")]))[i])]
new_ref <- cbind(new_ref, rowSums(m)/ncol(m))
}
colnames(new_ref) <- as.character(unique(rc.integrated$cb.anno.t[which(rc.integrated$cb.anno.t!="Unknown")]))
gene.var <- apply(new_ref,1,var)

# Select top 200 highly variable genes
new_ref_sub <- new_ref[names(sort(gene.var,decreasing = TRUE)[1:200]),]

# Merge 55 newly-built reference with atlas

merge.rownames <- function (x,y){
  dat <- merge(x = x, y = y, by = "row.names")
  rownames(dat) <- dat$Row.names
  dat <- dat[,-1]
  return(dat)
}
nr <- Reduce(merge.rownames, list(new_ref_sub,afm)) 
nr <- as.matrix(nr)

colnames(nr)[1:56]

1. 'Elongation-Atrichoblast-G2'
2. 'Elongation-Cortex-G2'
3. 'Meristem-Lateral Root Cap'
4. 'Elongation-Pericycle-G1'
5. 'Elongation-Endodermis-G2'
6. 'Elongation-Atrichoblast-G1'
7. 'Elongation-Phloem-G2'
8. 'Meristem-Quiescent Center'
9. 'Maturation-Cortex-G2'
10. 'Meristem-Atrichoblast-G2'
11. 'Elongation-Pericycle-G2'
12. 'Maturation-Trichoblast-G1'
13. 'Maturation-Atrichoblast-G1'
14. 'Maturation-Cortex-G1'
15. 'Meristem-Phloem-G2'
16. 'Elongation-Procambium-G1'
17. 'Maturation-Procambium-G1'
18. 'Maturation-Atrichoblast-G2'
19. 'Maturation-Columella'
20. 'Meristem-Atrichoblast-G1'
21. 'Maturation-Lateral Root Cap'
22. 'Elongation-Xylem-G1'
23. 'Elongation-Cortex-G1'
24. 'Maturation-Trichoblast-G2'
25. 'Elongation-Endodermis-G1'
26. 'Meristem-Columella'
27. 'Elongation-Phloem-G1'
28. 'Elongation-Trichoblast-G1'
29. 'Maturation-Xylem-G2'
30. 'Elongation-Procambium-G2'
31. 'Maturation-Pericycle-G1'
32. 'Maturation-Endodermis-G2'
33. 'Elongation-Trichoblast-G2'
34. 'Maturation-Xylem-G1'
35. 'Maturation-Pericycle-G2'
36. 'Elongation-Lateral Root Cap'
37. 'Meristem-Trichoblast-G1'
38. 'Elongation-Columella'
39. 'Maturation-Procambium-G2'
40. 'Elongation-Xylem-G2'
41. 'Maturation-Endodermis-G1'
42. 'Meristem-Cortex-G1'
43. 'Meristem-Trichoblast-G2'
44. 'Meristem-Pericycle-G2'
45. 'Meristem-Procambium-G2'
46. 'Meristem-Pericycle-G1'
47. 'Meristem-Procambium-G1'
48. 'Maturation-Phloem-G2'
49. 'Maturation-Phloem-G1'
50. 'Meristem-Cortex-G2'
51. 'Meristem-Endodermis-G1'
52. 'Meristem-Xylem-G1'
53. 'Meristem-Phloem-G1'
54. 'Meristem-Xylem-G2'
55. 'Meristem-Endodermis-G2'
56. 'AAACCCAAGCCATTCA_1'

# Correlation-based annotation using 55 newly-built references
nr_label=colnames(new_ref)
nr_stat <- suppressWarnings(sapply(56:ncol(nr), function(i) sapply(1:55, function(j) cor.test(nr[,i],nr[,j],method = "pearson")[c(3,4)])))
nr_cor <- nr_stat[seq(2,nrow(nr_stat),2),]
nr_pvalue <- nr_stat[seq(1,nrow(nr_stat)-1,2),]

# Update timezone-celltype annotation on second consensus annotation 
final.anno.P <- rep("Unknown",ncol(rc.integrated))
final.anno.cor.P <- rep(0,ncol(rc.integrated))
ct.list <- c("Quiescent Center", "Columella", "Lateral Root Cap", "Atrichoblast", "Trichoblast", "Cortex", "Endodermis", "Pericycle", "Phloem", "Xylem", "Procambium")

for (j in ct.list){
    tm.idx <- grep(j,nr_label)
    ct.idx <- grep(j,rc.integrated$cb.anno.2)
    nr_max <- sapply(ct.idx, function(i) max(as.numeric(nr_cor[tm.idx,i])))
    nr_ident <- sapply(ct.idx, function(i) nr_label[tm.idx][which(as.numeric(nr_cor[tm.idx,i])==max(as.numeric(nr_cor[tm.idx,i])))])
    names(nr_max) <- nr_ident
    final.anno.P[ct.idx] <- as.character(nr_ident)
    final.anno.cor.P[ct.idx] <- as.numeric(nr_max)
}

# Store updated timezone-celltype annotation 
rc.integrated$final.anno.P <- final.anno.P
rc.integrated$final.anno.cor.P <- final.anno.cor.P

table(rc.integrated$cb.anno.2.time)

Elongation Maturation   Meristem    Unknown 
     44681      28597      23705      13444 

# Plot timezone annotation of second consensus annotation
rc.integrated$cb.anno.2.time <- gsub("-.*$","",gsub("-.*-G2$","",gsub("-.*-G1$","",as.character(rc.integrated$final.anno.P))))
options(repr.plot.width=12, repr.plot.height=10)
DimPlot(rc.integrated, reduction = "umap", group.by = "cb.anno.2.time", order = c("Unknown","Maturation","Elongation","Meristem"),cols = c("#DCEDC8", "#42B3D5", "#1A237E", "#EEEEEE"))

11. Final annotation

# Extract integrated (batch-corrected) expression matrix
afm <- rc.integrated@assays$integrated@data

# Pool (average) expression values of each grouping
new_ref <- matrix(nrow=nrow(afm), ncol = 0)
for (i in 1:55) {
m <- afm[,which(rc.integrated$final.anno.P==as.character(unique(rc.integrated$final.anno.P[which(rc.integrated$final.anno.P!="Unknown")]))[i])]
new_ref <- cbind(new_ref, rowSums(m)/ncol(m))
}
colnames(new_ref) <- as.character(unique(rc.integrated$final.anno.P[which(rc.integrated$final.anno.P!="Unknown")]))
gene.var <- apply(new_ref,1,var)

# Select top 200 highly variable genes
new_ref_sub <- new_ref[names(sort(gene.var,decreasing = TRUE)[1:200]),]

# Merge 55 newly-built reference with atlas

merge.rownames <- function (x,y){
  dat <- merge(x = x, y = y, by = "row.names")
  rownames(dat) <- dat$Row.names
  dat <- dat[,-1]
  return(dat)
}
nr <- Reduce(merge.rownames, list(new_ref_sub,afm)) 
nr <- as.matrix(nr)

str(nr)
colnames(nr)[1:56]

 num [1:200, 1:110482] -1.8187 -0.0305 -1.805 -0.873 -1.1245 ...
 - attr(*, "dimnames")=List of 2
  ..$ : chr [1:200] "AT1G02900" "AT1G03870" "AT1G05260" "AT1G07640" ...
  ..$ : chr [1:110482] "Meristem-Atrichoblast-G2" "Elongation-Atrichoblast-G1" "Elongation-Cortex-G2" "Elongation-Columella" ...

1. 'Meristem-Atrichoblast-G2'
2. 'Elongation-Atrichoblast-G1'
3. 'Elongation-Cortex-G2'
4. 'Elongation-Columella'
5. 'Maturation-Columella'
6. 'Meristem-Lateral Root Cap'
7. 'Elongation-Xylem-G2'
8. 'Elongation-Lateral Root Cap'
9. 'Maturation-Pericycle-G2'
10. 'Elongation-Endodermis-G2'
11. 'Elongation-Atrichoblast-G2'
12. 'Elongation-Phloem-G2'
13. 'Maturation-Procambium-G1'
14. 'Meristem-Quiescent Center'
15. 'Elongation-Pericycle-G2'
16. 'Maturation-Trichoblast-G1'
17. 'Maturation-Atrichoblast-G1'
18. 'Maturation-Pericycle-G1'
19. 'Elongation-Pericycle-G1'
20. 'Maturation-Cortex-G2'
21. 'Meristem-Phloem-G1'
22. 'Elongation-Procambium-G1'
23. 'Meristem-Endodermis-G2'
24. 'Maturation-Xylem-G2'
25. 'Elongation-Trichoblast-G1'
26. 'Maturation-Procambium-G2'
27. 'Meristem-Atrichoblast-G1'
28. 'Elongation-Endodermis-G1'
29. 'Elongation-Cortex-G1'
30. 'Maturation-Lateral Root Cap'
31. 'Maturation-Cortex-G1'
32. 'Maturation-Atrichoblast-G2'
33. 'Elongation-Xylem-G1'
34. 'Meristem-Columella'
35. 'Maturation-Endodermis-G1'
36. 'Elongation-Phloem-G1'
37. 'Meristem-Phloem-G2'
38. 'Meristem-Procambium-G2'
39. 'Maturation-Trichoblast-G2'
40. 'Elongation-Trichoblast-G2'
41. 'Elongation-Procambium-G2'
42. 'Maturation-Endodermis-G2'
43. 'Meristem-Pericycle-G2'
44. 'Maturation-Phloem-G1'
45. 'Meristem-Trichoblast-G2'
46. 'Meristem-Cortex-G2'
47. 'Maturation-Phloem-G2'
48. 'Maturation-Xylem-G1'
49. 'Meristem-Trichoblast-G1'
50. 'Meristem-Xylem-G2'
51. 'Meristem-Pericycle-G1'
52. 'Meristem-Cortex-G1'
53. 'Meristem-Xylem-G1'
54. 'Meristem-Endodermis-G1'
55. 'Meristem-Procambium-G1'
56. 'AAACCCAAGCCATTCA_1'

# Correlation-based annotation using 55 newly-built references
nr_label=colnames(new_ref)
nr_stat <- suppressWarnings(sapply(56:ncol(nr), function(i) sapply(1:55, function(j) cor.test(nr[,i],nr[,j],method = "pearson")[c(3,4)])))
nr_cor <- nr_stat[seq(2,nrow(nr_stat),2),]
nr_pvalue <- nr_stat[seq(1,nrow(nr_stat)-1,2),]                                                                                                        
nr_max <- sapply(1:(ncol(nr)-55), function(i) max(as.numeric(nr_cor[,i])))
nr_ident <- sapply(1:(ncol(nr)-55), function(i) nr_label[which(as.numeric(nr_cor[,i])==max(as.numeric(nr_cor[,i])))])
nr_maxp <- sapply(1:(ncol(nr)-55), function(i) as.numeric(nr_pvalue[,i])[which(as.numeric(nr_cor[,i])==max(as.numeric(nr_cor[,i])))])                     
names(nr_max) <- nr_ident
rc.integrated$final.anno.P <- as.character(nr_ident)
rc.integrated$final.anno.cor.P <- as.numeric(nr_max)
rc.integrated$final.anno.pvalue.P <- as.numeric(nr_maxp)

# Plot new correlation-based annotation
rc.integrated$final.anno.wo.unknown <-  gsub("-G1$","",gsub("-G2$","",gsub("^Meristem-","",gsub("^Elongation-","",gsub("^Maturation-","",as.character(rc.integrated$final.anno.P))))))
rc.integrated$time.anno <- gsub("-.*$","",as.character(rc.integrated$final.anno.P))
order <- c("Quiescent Center", "Stem Cell Niche", "Columella", "Lateral Root Cap", "Atrichoblast", "Trichoblast", "Cortex", "Endodermis", "Pericycle", "Phloem", "Xylem", "Procambium", "Unknown")
palette <- c("#9400D3","#dcd0ff", "#5ab953", "#bfef45", "#008080", "#21B6A8", "#82b6ff", "#0000FF","#ff9900","#e6194b", "#9a6324", "#ffe119","#EEEEEE")
rc.integrated$final.anno.wo.unknown <- factor(rc.integrated$final.anno.wo.unknown, levels = order[sort(match(unique(rc.integrated$final.anno.wo.unknown),order))]) 
color <- palette[sort(match(unique(rc.integrated$final.anno.wo.unknown),order))]
options(repr.plot.width=12, repr.plot.height=10)
DimPlot(rc.integrated, reduction = "umap", group.by = "final.anno.wo.unknown", cols = color)+ggtitle("final annotation")
DimPlot(rc.integrated, reduction = "umap", group.by = "time.anno", order = c("Unknown","Maturation","Elongation","Meristem"),cols = c("#DCEDC8", "#42B3D5", "#1A237E", "#EEEEEE"))+ggtitle("timezone annotation")

12. Final touch (manually correct/adjust annotation)

# Extract annotations
tc.anno <- gsub("-G1$","",gsub("-G2$","",rc.integrated$final.anno.P))
tm.anno <- gsub("-.*$","",tc.anno)
ct.anno <- gsub("^Meristem-","",gsub("^Elongation-","",gsub("^Maturation-","",tc.anno)))

tc.anno <- paste(tm.anno, ct.anno, sep="-")
tc.anno[which(tc.anno=="Elongation-Putative Quiescent Center")]="Meristem-Putative Quiescent Center";
tc.anno[which(tc.anno=="Maturation-Putative Quiescent Center")]="Meristem-Putative Quiescent Center";
tc.anno[which(tc.anno=="Maturation-Stem Cell Niche")]="Meristem-Stem Cell Niche";
tc.anno[which(tc.anno=="Elongation-Stem Cell Niche")]="Meristem-Stem Cell Niche";

tm.anno <- gsub("-.*$","",tc.anno)
ct.anno <- gsub("^Meristem-","",gsub("^Elongation-","",gsub("^Maturation-","",tc.anno)))

# Annotation for Columella, LRC and Stem Cell Niche relies solely on marker-based annotation
LRC.idx <- which(rc.integrated$escore.anno=="Lateral Root Cap");
CLM.idx <- which(rc.integrated$escore.anno=="Columella");
ct.anno[which(ct.anno=="Quiescent Center")]="Stem Cell Niche"
ct.anno[LRC.idx] = "Lateral Root Cap"
ct.anno[CLM.idx] = "Columella"
tmp <- paste(tm.anno, ct.anno, sep="-")
SCN.idx <- which(rc.integrated$escore.anno=="Quiescent Center" & tmp!="Maturation-Atrichoblast");
ct.anno[SCN.idx] = "Stem Cell Niche"
ct.anno[QC.idx] = "Putative Quiescent Center" 
tc.anno <- paste(tm.anno, ct.anno, sep="-")

# Store final annotation
rc.integrated$celltype.anno <- ct.anno
rc.integrated$time.anno <- tm.anno
rc.integrated$time.celltype.anno <- tc.anno

# Plot final cell type annotation
order <- c("Putative Quiescent Center", "Stem Cell Niche", "Columella", "Lateral Root Cap", "Atrichoblast", "Trichoblast", "Cortex", "Endodermis", "Pericycle", "Phloem", "Xylem", "Procambium", "Unknown")
palette <- c("#9400D3","#dcd0ff", "#5ab953", "#bfef45", "#008080", "#21B6A8", "#82b6ff", "#0000FF","#ff9900","#e6194b", "#9a6324", "#ffe119","#EEEEEE")
rc.integrated$celltype.anno <- factor(rc.integrated$celltype.anno, levels = order[sort(match(unique(rc.integrated$celltype.anno),order))]) 
color <- palette[sort(match(unique(rc.integrated$celltype.anno),order))]
options(repr.plot.width=12, repr.plot.height=10)
DimPlot(rc.integrated, reduction = "umap", group.by = "celltype.anno", cols = color)+ggtitle("celltype annotation")

# Plot final time zone annotation
DimPlot(rc.integrated, reduction = "umap", group.by = "time.anno", order = c("Unknown","Maturation","Elongation","Meristem"),cols = c("#DCEDC8", "#42B3D5", "#1A237E", "#EEEEEE"))+ggtitle("timezone annotation")

# Save Seurat object
saveRDS(rc.integrated, file='./Root_Atlas.rds')